It has been well-established that a critical key to the long-term success of fuel cell vehicles is the development of a hydrogen infrastructure. Fuel cell vehicles are projected by many to be the eventual replacement of, or at the very least supplement to, the internal combustion engine vehicle. This is driven primarily by the growing concerns over greenhouse gas and air pollutant emissions, long-term availability of fossil fuels, and energy supply security. The proton exchange membrane (PEM) fuel cells which are the focus of almost all the current efforts towards the development of commercially viable fuel cell vehicles require hydrogen as a fuel. Virtually all efforts towards the on-board production of hydrogen from more portable hydrocarbon fuels have been abandoned in recent years, and almost all fuel cell vehicle manufacturers are currently focusing on refueling the vehicles with high-purity liquid or gaseous hydrogen.
The means by which hydrogen can be produced in large quantities are well understood. Steam reforming of methane is the primary means by which hydrogen is currently being produced on an industrial scale. Today, about half of the world production of hydrogen is used in oil refineries, mainly for the production of automotive fuels. Another 40% is consumed in the commercial production of ammonia. However, the annual production volume of hydrogen in the United States is comparable to only two days worth of gasoline consumption. Furthermore, hydrogen is currently being produced predominantly on a large industrial scale. For a successful transportation infrastructure, the hydrogen refueling network must be well-distributed. Hydrogen is, however, very problematic to distribute. Gaseous hydrogen has one of the lowest energy densities, making it difficult to transport in the amounts that would be required for a transportation fuel cell infrastructure. Distributing hydrogen in liquid form is also difficult—it requires very low temperatures (22 K), and even in liquid form hydrogen has a low energy density. Because of these concerns, it can be reasonably concluded that a hydrogen infrastructure capable of supplying the refueling needs of fuel cell vehicles will need to rely on the distributed production of high-purity hydrogen.
The widely distributed hydrogen production necessary for a transportation fuel cell infrastructure is much smaller than the typical refinery or ammonia-producing hydrogen production scale. Various means by which high-purity hydrogen can be economically produced at this small scale are currently being pursued. One such means of production is to apply, on a smaller scale, the well-understood methods of producing hydrogen at the current large scales. The predominant method of producing hydrogen at large scales is by steam reforming natural gas (methane) over a catalyst. The steam reforming reaction produces hydrogen and carbon monoxide as follows:CH4+H2O→3H2+CO
The steam reforming reaction is highly endothermic, requiring 206 kJ of energy per mole of methane consumed. Some of the CO produced is converted to CO2 via the associated water-gas shift reaction:CO+H2O CO2+H2 
This reaction is exothermic, and liberates 41 kJ of energy per mole of CO consumed. Steam reforming of methane is typically carried out at temperatures in the range of 700° C.-900° C. Since the reaction is endothermic, heat must be supplied to the reactor. This is typically accomplished by loading the catalyst into a series of tubes which are placed in a furnace. The hydrogen can be extracted from the steam reforming product gas (reformate) through various well-understood means, such as metal membrane or pressure swing adsorption (PSA).
It has long been understood that in order to make steam reforming of natural gas feasible at the smaller scales required for a distributed production of hydrogen for fuel cell vehicles, a greater integration between the heat-producing combustor and the endothermic steam reforming reaction is needed. Attempts to build such systems have met with some success in the past, but the performance efficiency has always been limited by the ability to transfer the required heat into the steam reforming reaction without generating extremely high (>1000° C.) metal temperatures.